25
20
1719708
170227
So excited this book finally got delivered today! One of my big goals for 2022 is to learn Bayesian stats. I’m excited to learn! If you have any other resource suggestions, please let me know! #rstats #bayes pic.twitter.com/TQB7Fu8XGo
— Micah Hirsch, M.S. (@MicahEHirsch) January 7, 2022
Best of Nmap Cheatsheet https://t.co/MQ7U0sw4HL #BigData #Analytics #DataScience #AI #MachineLearning #IIoT #PyTorch #Python #RStats #TensorFlow #Java #JavaScript #ReactJS #GoLang #CloudComputing #Serverless #DataScientist #Linux #Programming #Coding #100DaysofCode #DL #NLProc pic.twitter.com/KaUOEBXCtf
— Syeda Sheraj Ali (@Sheraj99) January 7, 2022
A Comprehensive Guide of Regularization. #BigData #Analytics #DataScience #AI #MachineLearning #IoT #IIoT #Python #RStats #TensorFlow #Java #JavaScript #ReactJS #GoLang #CloudComputing #Serverless #DataScientist #Linux #Programming #Coding #100DaysofCode https://t.co/kgFHACtL9W pic.twitter.com/5tlVvXncaS
— Dr. Ganapathi Pulipaka 🇺🇸 (@gp_pulipaka) January 7, 2022
| User | Engagement/Tweet |
|---|---|
| @CloarecJulien | 7589.182 |
| @v_matzek | 2453.000 |
| @kaymwilliamson | 1864.000 |
| @TheToadLady | 1602.500 |
| @kiramhoffman | 1138.000 |
| @OwenOzier | 959.000 |
| @Kunkakom | 923.000 |
| @drhammed | 892.000 |
| @SebastienPolis | 875.000 |
| @rpydaneogrendim | 765.000 |
Where Engagement is RT * 2 + Favourite
Relationships in the graph describe replies and quote retweets from the top tweeters that also have the hashtag.
---
title: "#rstats Twitter Explorer"
output:
flexdashboard::flex_dashboard:
orientation: rows
vertical_layout: scroll
source_code: embed
theme:
version: 4
bootswatch: yeti
css: styles/main.css
---
```{r load_proj, include=FALSE}
devtools::load_all()
```
```{r load_packages, include=FALSE, cache=TRUE}
library(flexdashboard)
library(rtweet)
library(dplyr)
library(stringr)
library(tidytext)
library(lubridate)
library(echarts4r)
library(DT)
rstats_tweets <- read_twitter_csv("data/rstats_tweets.csv.gz") %>%
mutate(created_at = as_datetime(created_at))
```
```{r time_data, include=FALSE, cache=TRUE}
count_timeseries <- rstats_tweets %>%
ts_data(by = "hours")
tweets_week <- rstats_tweets %>%
filter(date(created_at) %within% interval(floor_date(today(), "week"), today()))
tweets_today <- rstats_tweets %>%
filter(date(created_at) == today())
```
```{r numbers, include=FALSE, cache=TRUE}
number_of_unique_tweets <- get_unique_value(rstats_tweets, text)
number_of_unique_tweets_today <-
get_unique_value(tweets_today, text)
number_of_tweeters_today <- get_unique_value(tweets_today, user_id)
number_of_likes <- rstats_tweets %>%
pull(favorite_count) %>%
sum()
```
```{r rankings_data, include=FALSE, cache=TRUE}
top_tweeters <- rstats_tweets %>%
group_by(user_id, screen_name, profile_url, profile_image_url) %>%
summarize(engagement = (sum(retweet_count) * 2 + sum(favorite_count)) / n()) %>%
ungroup() %>%
slice_max(engagement, n = 10, with_ties = FALSE)
top_tweeters_format <- top_tweeters %>%
mutate(
profile_url = stringr::str_glue("https://twitter.com/{screen_name}"),
screen_name = stringr::str_glue('@{screen_name}'),
engagement = formattable::color_bar("#a3c1e0", formattable::proportion)(engagement)
) %>%
select(screen_name, engagement)
top_hashtags <- rstats_tweets %>%
tidyr::separate_rows(hashtags, sep = " ") %>%
count(hashtags) %>%
filter(!(hashtags %in% c("rstats", "RStats"))) %>%
slice_max(n, n = 10, with_ties = FALSE) %>%
mutate(
number = formattable::color_bar("plum", formattable::proportion)(n),
hashtag = stringr::str_glue(
'#{hashtags}'
),
) %>%
select(hashtag, number)
word_banlist <- c("t.co", "https", "rstats")
top_words <- rstats_tweets %>%
select(text) %>%
unnest_tokens(word, text) %>%
anti_join(stop_words) %>%
filter(!(word %in% word_banlist)) %>%
filter(nchar(word) >= 4) %>%
count(word, sort = TRUE) %>%
slice_max(n, n = 10, with_ties = FALSE) %>%
select(word, n)
top_co_hashtags <- rstats_tweets %>%
unnest_tokens(bigram, hashtags, token = "ngrams", n = 2) %>%
tidyr::separate(bigram, c("word1", "word2"), sep = " ") %>%
filter(!word1 %in% c(stop_words$word, word_banlist)) %>%
filter(!word2 %in% c(stop_words$word, word_banlist)) %>%
count(word1, word2, sort = TRUE) %>%
filter(!is.na(word1) & !is.na(word2)) %>%
slice_max(n, n = 100, with_ties = FALSE)
top_locations <- rstats_tweets %>%
filter(!is.na(location) & location != "#rstats") %>%
distinct(user_id, .keep_all = TRUE) %>%
mutate(location = str_replace_all(location, "London$", "London, England")) %>%
count(location) %>%
slice_max(n, n = 10, with_ties = FALSE)
```
Home {data-icon="ion-home"}
====
Row
-----------------------------------------------------------------------
### Tweets Today
```{r tweets_today}
valueBox(number_of_unique_tweets_today, icon = "fa-comment-alt", color = "plum")
```
### Tweeters Today
```{r tweeters_today}
valueBox(number_of_tweeters_today, icon = "fa-user", color = "peachpuff")
```
### #rstats Likes
```{r likes}
valueBox(number_of_likes, icon = "fa-heart", color = "palevioletred")
```
### #rstats Tweets
```{r unique_tweets}
valueBox(number_of_unique_tweets, icon = "fa-comments", color = "mediumorchid")
```
Row {.tabset .tabset-fade}
-----------------------------------------------------------------------
### Tweet volume
```{r tweet_volume}
plot_tweet_volume(count_timeseries)
```
### Tweets by Hour of Day
```{r tweets_by_hour}
plot_tweet_by_hour(rstats_tweets)
```
Row
-----------------------------------------------------------------------
### 💗 Most Liked Tweet Today {.tweet-box}
```{r most_liked}
most_liked_url <- tweets_today %>%
slice_max(favorite_count, with_ties = FALSE)
get_tweet_embed(most_liked_url$screen_name, most_liked_url$status_id)
```
### ✨ Most Retweeted Tweet Today {.tweet-box}
```{r most_rt}
most_retweeted <- tweets_today %>%
slice_max(retweet_count, with_ties = FALSE)
get_tweet_embed(most_retweeted$screen_name, most_retweeted$status_id)
```
### 🎉 Most Recent {.tweet-box}
```{r most_recent}
most_recent <- tweets_today %>%
slice_max(created_at, with_ties=FALSE)
get_tweet_embed(most_recent$screen_name, most_recent$status_id)
```
Rankings {data-icon="ion-arrow-graph-up-right"}
=========
Row
-----------------------------------------------------------------------
### Top Tweeters
```{r top_tweeters}
top_tweeters_format %>%
knitr::kable(
format = "html",
escape = FALSE,
align = "cll",
col.names = c("User", "Engagement/Tweet "),
table.attr = 'class = "table"'
)
```
Where Engagement is `RT * 2 + Favourite`
### Network of top tweeters
Relationships in the graph describe replies and quote retweets from the top tweeters
that also have the hashtag.
```{r top_tweeters_net}
edgelist <-
network_data(rstats_tweets %>% unflatten(), "reply,quote")
nodelist <- attr(edgelist, "idsn") %>%
bind_cols()
top_edges <- edgelist %>%
filter((from %in% top_tweeters$user_id) |
(to %in% top_tweeters$user_id))
top_nodes <- nodelist %>%
filter((id %in% top_edges$from) | (id %in% top_edges$to)) %>%
mutate(is_top = ifelse((id %in% top_tweeters$user_id), "yes", "no"),
size = 10)
e_charts() %>%
e_graph() %>%
e_graph_nodes(top_nodes, id, sn, size, category = is_top, legend = FALSE) %>%
e_graph_edges(top_edges, from, to) %>%
e_tooltip()
```
Row
-----------------------------------------------------------------------
### Top Words
```{r top_words}
top_words %>%
e_charts(word) %>%
e_bar(n, legend = FALSE) %>%
e_x_axis(
axisLabel = list(
interval = 0L,
rotate = 30
)
) %>%
e_toolbox_feature("saveAsImage") %>%
e_axis_labels(y = "Number of occurrences")
```
### Top Locations
```{r top_locations}
top_locations %>%
mutate(location = str_wrap(location, 9)) %>%
e_charts(location) %>%
e_bar(n, legend = FALSE) %>%
e_x_axis(
axisLabel = list(
interval = 0L,
rotate = 30
)
) %>%
e_toolbox_feature("saveAsImage") %>%
e_axis_labels(y = "Number of users from location")
```
Row
-----------------------------------------------------------------------
### Top Hashtags
```{r top_hashtags}
top_hashtags %>%
knitr::kable(
format = "html",
escape = FALSE,
align = "cll",
col.names = c("Hashtag", "Count"),
table.attr = 'class = "table"'
)
```
Excluding `#rstats` and similar variations
### Common co-occuring hashtags
Hashtags that occur together, grouped by community detection
```{r co_hashtags}
top_co_hash_nodes <- tibble(
nodes = c(top_co_hashtags$word1, top_co_hashtags$word2)
) %>%
distinct()
e_chart() %>%
e_graph() %>%
e_graph_nodes(top_co_hash_nodes, nodes, nodes, nodes) %>%
e_graph_edges(top_co_hashtags, word1, word2) %>%
e_modularity()
```
Data {data-icon="ion-stats-bars"}
==============
### Tweets in the current week {.datatable-container}
```{r datatable}
tweets_week %>%
select(
status_url,
created_at,
screen_name,
text,
retweet_count,
favorite_count,
mentions_screen_name
) %>%
mutate(
status_url = stringr::str_glue("On Twitter")
) %>%
datatable(
.,
extensions = "Buttons",
rownames = FALSE,
escape = FALSE,
colnames = c("Timestamp", "User", "Tweet", "RT", "Fav", "Mentioned"),
filter = 'top',
options = list(
columnDefs = list(list(
targets = 0, searchable = FALSE
)),
lengthMenu = c(5, 10, 25, 50, 100),
pageLength = 10,
scrollY = 600,
scroller = TRUE,
dom = '<"d-flex justify-content-between"lBf>rtip',
buttons = list('copy', list(
extend = 'collection',
buttons = c('csv', 'excel'),
text = 'Download'
))
)
)
```